J . Org. Chem., Vol. 36, No. 9, 29YO 2877
SYNTHESIS OF p-AMINOBENZOYL PEPTIDES J4t15,= 4 Hz, C5rH2), 4.05 (hex, 1, J 3 t 1 4 t = 2.5 Ha, J 4 t . 6 , = 4 Ha, C4,H),4.54 (9, 1 , Jl,,zf = 8.5 Ha, J~t,3.= 6H2, Cz,H),4.96 (q, 1, J2,,3t = 6 Hz, J3t14t = 2.5 Ha, C3*H),5.88 (d, 1 , J 5 . a 8 Ha, CEH), 6.37 (d, 1, J I , , = ~ 8.5 Hz, CltH), 7.61 (d, 1, J 5 , 6 = 8 Ha, CaH), 9.81 (br s, 1, NH); mass spectrum (70 eV) m/e 506 (M+), 379 (M - I ) , 319 (M - I - AcOH), 192 (M - Iz - AcOH). Anal. Calcd for C11H12N20612: C, 26.11; H , 2.39; N, 5.54. Found: C, 26.04; H, 2.43; N, 5.08. 3'-O-Acetyl-2',5'-dideoxyuridine (34d).-A solution of 34c (85 mg) in 85% methanol (8 ml) containing sodium acetate (84 mg) was hydrogenated for 2 hr at 25" in the presence of 10% palladium on charcoal (32 me). The mixture was then filtered through Celite, evaporated, and partitioned between ethyl acetate and very dilute aqueous sodium thiosulfate. Evaporation of the organic phase left 21 mg (50%) of 34d as crystals, mp 182-183'. An analytical sample from chloroform-hexane had mp 185.5186":"::A: 260 mp (E 10,200); ORD (MeOH) positive Cotton effect with a peak at 283 mp (@ +4400°), crossover a t 273 mp and a trough at 254 mp (4, -10,000"); nmr of the crude or recrystallized sample (CDCls) was very sharp with 1.42 ppm (d, 3, J4,,5t = 6.5 Hz, C5tH3), 2.12 (s, 3, OAc),2.16 (oct, 1, Jgm = 15 3
Ha, J 1 t , 2 t a = 8 Ha, Jzta,a, = 6.5H2, CztSH), 2.54 (oct, 1, J g e m = 15 Ha, Jlt.z!b = 6 HE, J2'b,3' = 3 Ha, CzPbH), 4.23 (OCt, 1 , J4,,6,= 6.5 Ha, J3,,,, = 3 Hz, C4,H), 4.88 (quint, 1, JZ'b.3' = J 3 , , , = 3 Ha, J2,a.v = 6.5 Ha, CatH), 5.81 (d, 1, J s , ~= 8 Hz, CsH), 6.21 (9, 1 , Jl,,z,, = 8 Ha, J l t . 2 ' b = 6H2, CIJI), 7.46 (d, 1, J 5 , 6 = 8 Ha, C6H). Anal. Calcd for CllH14N206: C, 51.96; H, 5.55; N, 11.02. Found: C, 52.14; H , 5.96; N, 10.76.
Registry N0.-4a, 25442-40-4; 4b, 25442-42-6; Sa, 14260-81-2; 5by14259-59-7; SC,25442-44-8; 7a, 1426087-8; 7b, 14260-83-4; 9c, 25383-77-1 ; ~ O C 25442-45-7; , 17, 25383-78-2; 19by 25442-46-0; 21, 25383-79-3; 22, 25383-80-6; 27, 25383-81-7; 28a, 25442-47-1 ; 28b, 25383-82-8; 29,25442-48-2; 30,24514-27-0; 32, 2544249-3; 33a, 25383-84-0; 34c, 25383-85-1 ; 34d, 2544250-6; 3'-deoxy-3'-iodothymidine, 14260-82-3; methyltriphenoxyphosphonium iodide, 17579-99-6.
Synthesis of p-Aminobenzoyl Peptideslavb A. R. MITCHELL,'~ S. K. GUPTA,AND ROGERW. R O E S K E ' ~ ~ ~ Department of Biochemistry, Indiana University School of Medicine, Indianapolis, Indiana &3WOW Received December 17, 1969 Cyclic peptides incorporating p-aminobenzoic acid are proposed as enzyme models. The p-aminobenzoyl residues may provide a relatively apolar cavity and substrate binding site, and the peptide bridges joining the p-aminobenzoyl residues allow the placement of functional side chains which can serve as a catalytic site. The synthesis of glycyl-p-aminobenzoylglycyl-im-ben~yl-~-histidylglycyl-~-aminobenaoyl-~-aminocaproicacid dihydrobromide (4) was carried out using the solid-phase method of peptide synthesis. Peptide 4 was cyclized using excess N,N '-dicyclohexylcarbodiimide in aqueous methanol to give cyclo-(glycyl-p-aminobenaoylglycyl-im-benzylL-histidylglycyl-p-aminobenzoyl-e-aminocaproyl) (3). Peptide 3 was hydrogenated to give cyclo-(glycyl-pamino benzoylglycyl-~-histidylglycyl-p-aminobenzoyl-r-aminocaproyl) (2), a simple example of the proposed class of peptides. The peptide was not sufficiently soluble in water to test its validity as an enzyme model. The saponification of p-aminobenaoyl peptide esters proceeds without major side reactions, contrary to reports in the literature. p-Aminobenzoyl peptides are cleaved by sodium in liquid ammonia.
The use of cyclic molecules as enzyme models has been explored in recent years. Synthetic cyclic p e p t i d e P 5 and cycloamyloses6-9 have been investigated. We propose molecules of the type 1 as enzyme models. The incorporation of p-aminobenzoyl residues m CO)NH+CO(NH
/
RCH h)$O-@H,(CO
\
HCR
I
'
1 (1) (a) This work was supported in part by grants from the U. 8. Public Health Service (GM10591) and the National Science Foundation (GB6631). (b) This paper was reported in part a t the 158th National Meeting of the American Chemical Society, New York, N. Y., Sept 1969, Abstract ORGN 21. (0) Abstracted in part from a dissertation submitted by A. R . Mitchell to Indiana University in partial fulfillment of the requirements for the Ph.D. degree. Financial support from the U. S. Department of Health, Education. and Welfare is gratefully acknowledged. (d) To whom inquiries should be addressed. (e) Research Career Development Awardee of the U. S. Public Health Service. (2) K. D. Kopple and D. E. Nitecki, J . Amer. Chem. S O C .88, , 4103 (1961). (3) K. D. Kopple and D. E. Nitecki, ibid.,84, 4457 (1962). (4) K. D. Kopple, R. R. Jarabak, and P. L. Bhatia, Biochemistry, 2 , 958 (1962). (5) J. C. Sheehan and D. N . McGregor, J . Amer. Chem. SOC.,84, 3000 (1962). (6) R . L. Van Etten, J. F. Sebastian, G. A. Clowes, and M. L. Bender, i b i d . , 89, 3242 (1967). (7) R. L. Van Etten, G. A. Clowes, J. F. Sebastian, and M. L. Bender, i b i d . , 89, 3253 (1967). (8) F. Cramer and G. Mackensen, Angew. Chem., Int. E d . Engl., 5, 601 (1966). (9) F. Cramer and G. Hettler, Naturwissenschaften, 64, 625 (1967).
into a cyclic peptide provides a relatively apolar cavity that, in aqueous solution, might act as a substrate binding site. The peptide bridges between the p-aminobenzoyl residues allow the placement of functional side chains which can serve as a catalytic site. The preparation of p-aminobenzoyl peptides using conventional methods of peptide synthesis has received limited a t t e n t i ~ n , ' ~ -and ' ~ the solid-phase method has not been used at all. In this communication we report the synthesis of the cyclic heptapeptide 2, and also our investigation of two side reactions accompanying the synthesis of p-aminobenzoyl peptides. c H 2 c o ~ H c H , c o ~~ HC O ~ H C H 2 c O N H
I
I
2,R=H 3, R = C6H5CH2-
The synthesis of 2 is outlined in Figure 1. The linear heptapeptide 4 was prepared by the solid-phase method of Merrifield, l 4 starting with N-t-butyloxy(10) F. E. King, J. W. Clark-Lewis, D. A. A. Kidd, and G. R. Smith, J . Chem. Soc., 1039 (1954). (11) W. Langenbeck and D. Weisbrod, J . Prakt. C h e n . . a8, 78 (1965). (12) W. Langenbeck and D. Weisbrod, ibid., 81, 92 (1966). (13) R. Geiger, K. Sturm, and W. Siedel, Chem. Ber., 101, 1223 (1968). (14) J. M. Stewart and J. D. Young, "Solid Phase Peptide Syntheais," W. E.Freeman, San Francisco, Calif., 1969.
2878
J. Org. Chem., Vol. 55, No. 9, i970
hhTCHELL,
BOC-Eac-OCH2-P
genolysis an additional spot due to a monobenzyl derivative. Determination of the molecular weight of 2 by X-ray diffraction also indicated a monomer, although the imperfect form of the crystals did not allow’an accurate measurement (see Experimental Section). When peptide 4 was cyclized and hydrogenated without isolation of 3, peptide 2 was obtained in 23% yield after chromatography. Titration of 2 in dimethyl sulfoxidewater (2:l) indicated a pK’ of 5.8, compared with a value of 6.4 for imidazole in the same solvent. The cyclic peptide 2 has a very low solubility (2 X M ) in neutral and basic aqueous solutions, which precludes a detailed study of its interaction with substrates. However, we mere able to measure its activity in catalyzing the hydrolysis of an easily hydrolyzed substrate, 2,4-dinitrophenyl acetate. Table I summarizes the kinetic results. Peptide 2 is less active than unsubstituted imidazole in the catalysis of hydrolysis of 2,4-dinitrophenyl acetate, which is not surprising since molecular models indicate that the imidazole side chain of the peptide cannot interact easily with the substrate’s carbonyl group when the substrate is in the cavity of the peptide ring. We are currently working on the synthesis of cyclic 0-,m-, and p-aminobenzoyl peptides that are designed to be more water soluble than 2 and to have cooperatively interacting functional groups. It has been claimed12 that extensive peptide bond scission occurs during the basic hydrolysis of methyl, ethyl, and p-nitrophenyl esters of p-aminobenzoyl peptides. In order to evaluate this quantitatively, we synthesized N- t-butyloxycarbonylglycyl-p-aminobenzoylglycine methyl ester (6) and, after treating it with sodium hydroxide and removing the BOC group with trifluoroacetic acid, analyzed the products with an amino acid analyzer. The results are shown in Scheme I. Since no glycyl-p-aminobenzoic acid could be
1. CFsCOzH-CHzClz
2.
3.
(CzHa)aN-CHCla BOC-Gly-PBb, DCC
BOC-Gly-Pab-Eac-OCHz-P
Bz
I
BOC-Gly-Pab-Gly-L-His-Gly-Pa b-Eac-OCHz-P JHBr-CFaCOOH
Bz
I
Gly-Pab-Gly-L-His-Gly-Pab-Eac .2HBr 4 I D C C , CHaOH-Hz0
Bz cyclo-(Gly-Pab-Gly-L-H!s-Gl y-Pab-Eac j 3
J
HpPd
Figure 1.-Outline of synthesis of cyclo-(glycyl-p-aminobenzoylglycy1-L- histidylglycyl- p aminobenzoyl- e-aminocaproyl) (2); Eac, e-aminocaproyl; Pab, p-rtminobenzoyl; BOC-, tbutyloxycarbonyl-; Bz, benzyl; DCC, N,N’-dicyclohexylcarbodiimide; P, polystyrene-2% divinyl benzene copolymer.
-
carbonyl-eaminocaproyl resin. l5 The peptide chain was lengthened using 4 equiv of the t-butyloxycarbonyl derivatives of glycyl-p-aminobenzoic acid, irn-benzylL-histidine, glycine, and glycyl-p-aminobenzoic acid, in that order. The dipeptide, t-butyloxycarbonylglycylp-aminobenzoic acid ( 5 ) , was added as a unit because we were unable to prepare t-butyloxycarbonyl-paminobenzoic acid conveniently. Addition of the dipeptide as a unit also enabled us to monitor the coupling a t this step with the amino acid analyzer, which we could not have done otherwise, since p-aminobenzoic acid does not give a positive ninhydrin test. Compound 5 was prepared from t-butyloxycarbonylglycine and p-aminobenzoic acid by the mixed anhydride procedure. All coupling reactions were carried out using N,N’-dicyclohexylcarbodiimide and were allowed to proceed for at least 15 hr. The linear heptapeptide was cleaved from the resin with hydrogen bromide in trifluoroacetic acid and purified by chromatography on Sephadex LH-20. The yield of purified 4 was 33%, based on e-aminocaproyl resin. Peptide 4 was cyclized using a tenfold excess of N,N‘-dicyclohexylcarbodiimide in aqueous methanol16 to yield protected cyclic peptide 3 in 25% yield. Debenzylation of 3 by catalytic hydrogenolysis was monitored by thin layer chromatography. The slow disappearance of 3 and the appearance of a single ninhydrin-negative and Paulypositive spot is evidence for the monomeric structure of 3 and 2. Had 3 been a dimer containing two histidine residues, we should have observed during the hydro(15) M. Rothe and H. Schneider, Angew. Chem., Int. Ed. Bngl., 5, 417 (1966). (16) T. Wieland and K. W. Ohly, Justus Liebige Ann. Chem., 605, 179 (1957).
SCHEME
GUPTA,AND ROESKE
1
BOC-NHCH&ONH 6
1
1 NaOH ( 2 equiv), MeOH-”0, 2 CF,COOH, 26”, 30 min
NH,CH&OOH
+
25-’, 15 min
NHz~ C O N H C H ~ C O , H
+ 7, al%
detected in the products, we assume that all the glycine arose from hydrolysis of the peptide bond preceding the p-aminobenzoic residue. When the saponification was allowed to run for 29 hr, glycine was formed in 8.4% yield and 7 in 92y0 yield. I n a control experiment in which 6 was treated directly with trifluoroacetic acid at 2.5’ for 30 min, no glycine was formed. flubsequently, numerous hydrolytic reactions have been carried out on esters of other peptides containing p-aminobenzoyl residues, and good yields of the intact
SYNTHESIS OF ~AMINOBENZOYL PEPTIDES
J. Org. Chem., Vol. 36, N o . 9, 1970
TABLE I HYDROLYSIS O F 2,4-DINITROPHENYL ACETATE (6 X lo-' M ) I N 0.01 M PHOSPHATE BUFFER(PH 7.19) AT 26' Catalyst
10akobsdlmin-l (%)
Molarity
Determinations
2879
a
kr, l./mol/minb
Buffer
3.58 f 0.12 (3.4) 6 1 . 0 x 10-6 3.75 f 0 . 0 2 (0.5) 3 170 2 . 0 x 10-6 3.89 f 0.02 (0.5) 3 155 Imidazole 1 . 0 x 10-6 3.90 f 0.05 (1.3) 3 320 Imidazole 2 . 0 x 10-6 4.15 f 0.12 (2.9) 3 300 a The solutions were 0.2% in acetonitrile and 0-0.15% in dimethylformamide. kz = (kobsd - k8,,~v)/c where kobsd is the observed first-order rate constant measured in the presence of catalyst, ksolvis the constant in the absence of peptide or imidazole catalysts, and c is the molar concentration of catalyst. 2 2
peptide carboxylic acid have been obtained." Compound 6 was prepared by coupling t-butyloxycarbonylglycine and p-aminobenzoylglycine methyl ester18 using 1-ethyl-3- (N,N-dimethylaminopropyl)carbodiimide hydrochloride.lg The treatment of p-aminobenzoyl peptides with sodium in liquid ammonia, e.g., to remove the iminobenzyl protecting group of histidine, results in extensive fragmentation of the peptide chain. Peptide 7 was almost completely destroyed under these conditions, and some of the products of the reaction were identified (Scheme 11). Reductive cleavage of the p-aminobenzoyl SCHEME I1
m
NH,CH,CONH
\ CONHCH,COOH -(__t7
1
Na / NH3 (-33O, 30 sec)
7
0.9%
+
NH,CH,COOH 37%
+
I-\
NH,CH,CONH~CH,OH 8,19%
residue was suggested by analogy to the cleavage of benzoyl amino acids by tetramethylammonium, generated a t a mercury electrodelZ0but only partially accounts for the degradation of 7 by sodium in liquid ammonia. The product 8 was synthesized by coupling N-carbobenzyloxyglycine with p-aminobenzyl alcoholz1 using DCC, followed by catalytic hydrogenolysis of the carbobenzyloxy group. The partial destruction of an N-p-aminobenzoyl derivative of lysine vasopressin by sodium in liquid ammonia was recently reported,13 and the reductive cleavage of acylproline bonds by the same reagent has been observed.zz-24
Experimental Section26 N-t-Butyloxycarbonylglycyl-p-aminobenzoic Acid (5).-A solution of N-t-butyloxycarbonylglycine26(14.0 g, 0.080 mol) and tri(17) Experiments by Dr. Anand 8. Dutta in this laboratory, (18) K. Miyatake and 8. Kaga, Japan Patent 6468 (1951); Chem. Abstr., 47,11251a (1953). (la) J. C . Sheehan, P. A. Cruickshank, and G. L. Boshart, J . Ore. Chem., 16, 2525 (1961). (20) L. Horner and H. Neumann, Chem. Ber., 98,3462 (1965). (21) 0. Fischer and G.Fischer, i b i d . , 18, 879 (1895). (22) W. F. Beniaek and R . D. Cole, Biochem. Biophys. Res. Comm., 10, 656 (1965). (23) W. F. Benisek, M. A. Raftery, and R. D. Cole, Biochemistry, 6, 3780 (1967). (24) M. Wilchek, S. Sarid, and A. Patchornik, Biochim. Biophys. Acta, 104,616 (1965). (26) Melting points were taken in capillary tubas and are corrected. Elemental analyses were performed by Midwest Microlab, Indianapolis,
ethylamine (11.2 ml, 0.081 mol) in 200 ml of tetrahydrofuran was cooled to -15' in an ice-salt bath. Isobutyl chloroformate (10.2 ml, 0.078 mol) was added over a 10-min period. p-Aminobenzoic acid (11.0 g, 0.080 mol) in 50 ml of tetrahydrofuran was added. The reaction mixture was stirred in the ice bath for 1hr, followed by stirring a t room temperature for 10 hr. The mixture was evaporated in uucuo to a moist residue which was dissolved in 100 ml of 50% aqueous acetic acid. The solution was chilled and 600 ml of water was added with vigorous stirring. An offwhite material (18.1 g) precipitated, mp 153-157'. The crude product was crystallized from tetrahydrofuran-petroleum ether (bp 30-60), yielding 10.4 g (44%) of a white powder, mp 167169', Rr (TCW) 0.50. Recrystallization from acetone gave an analytical sample, mp 170-172'. Anal. Calcd for CI4HH1gN2O6: C, 57.13; H , 6.16; N , 9.52. Found: 0 , 57.35; H , 6.44; N, 9.57. N-t-Butyloxycarbonylglycyl-p-aminobenzoylglycineMethyl Ester (6).-A solution of N-t-butyloxycarbonylglycine (14.7 g, 0.084 mol) and p-aminobenzoylglycine methyl esteP (15.6 g 0.075 mol) in dichloromethane was cooled in an ice bath and 1 -ethyl- 3 - (N ,N-dimethylaminopropyl)carbodiimide hydrochloride19 (15.0 g, 0.074 mol) was added. This was stirred at icebath temperature for 1 hr and overnight a t room temperature. The resulting solution was washed with water, 1% sodium bicarbonate, 1% citric acid, and water. After the solution was dried over magnesium sulfate, the methylene chloride was removed in vacuo to give a white material which was crystallized from methanol-water (1: 1) to give 24.0 g (87%) of white plates, 173-174', Rr (CMA) 0.56. An analytical sample was recrystallized from ethyl acetate, mp 167.5-168'. Anal. Calcd for C1,Hz3N306: C, 55.90; H, 6.31; N, 11.50. Found: C, 56.29; H, 6.57; N, 11.40. N-t-butyloxy carbonylglycyl-p-aminobenzaylglycine.-Compound 6 (1.30 g, 0.0037 mol) was saponified for 30 min in a solution containing 5 ml of 1.5 N sodium hydroxide and 15 ml of methanol. The light yellow solution obtained was diluted with water (50 ml) and extracted with ethyl acetate. The ethyl acetate extract was discarded and the aqueous phase was acidified to pH 3.0 with citric acid, saturated with sodium chloride, and extracted with ethyl acetate. The ethyl acetate extract was washed with water, dried over magnesium sulfate, and evaporated in vacuo to give 1.00 g (96.1%) of a white crystalline compound: decomposing over 320'; RI (BAWP) 0.59, Rr (PW) 0.56. The analytical sample was obtained through recrystallization from methanol-ethyl acetate-ether. Anal. Calcd for CleHzlNaO~:C, 54.69; H, 6.02; N, 12.01. Found: C, 54.73; H,6.20; N, 12.18. Glycyl-p-aminobenzoylglycine (7).-The above compound (0.20 g, 0.57 mmol) was treated with trifluoroacetic acid (2 ml) for 1 hr a t 25". The excess solvent was removed in vucuo and the resulting residue was triturated with ethyl acetate to give 200 mg of a white solid, melting and decomposing over 215'. The compound was crystallized from methanol-water-ethyl acetate (5: 1 :10) to yield the free tripeptide, melting with decomposition slowly over 215O, Rt (BAWP) 0.28. Ind. The solvents used for thin layer chromatography on silica gel G were n-butyl alcohol-acetic acid-water (BAW) in a ratio of 4: 1 : 1 , n-butyl alcoholacetic acid-water-pyridine (BBWP) in a ratio of 30: 6 : 24: 20, chloroformmethanol-acetic acid (CMA) in a ratio of 85:10:5, propanol-water (PW) in a ratio of 70:30, and tetrahydrofuran-cyclohexane-water (TCW) in a ratio of 93:7:5. A Technicon amino acid autoanalyzer waa used for amino acid and peptide determinations. Amino acid hydrolyses were carried out in 6 N HC1 in sealed evacuated tubes for 20 hr at 110° unless otherwise specified. (26) E.Schnabel, Justus Liebigs Ann. Chem., 702, 188 (1967).
2880 J . Org. Chem., Vol. 36, N o . 9, 19YO
MITCHELL, GUPTA,AND ROESKE
Anal. Calcd for C11H14NaOl: C, 52.58; H, 5.21; N , 16.72. duced 9.24 mmol of N,N'-dicyclohexylcarbodiimide in 10 ml of Found: C, 52.35; H , 5.36; N, 16.46. coupling solvent and let reaction proceed for 15 hr, (11) washed N-Carbo benzoxyglycyl-p-aminobenzyl Alcohol .-N-Carbo(three 50-ml portions) with coupling solvent, and (12) washed benzoxyglycine (2.09 g, 0.010 mol) and p-aminobenzyl alcohol21 (three 50-ml portions) with ethanol. (1.23 g, 0.010 mol) were dissolved in 20 ml of tetrahydrofuran The coupling solvents for the incorporation of N-t-butyloxyand cooled in an ice bath. N,N'-Dicyclohexylcarbodiimide (2.27 carbonylglycyl-p-aminobenzoicacid (S), N-t-butyloxycarbonylg, 0.011 mol) was added in 5 ml of tetrahydrofuran. The soluim-benzyl-L-histidine, and N-t-butyloxycarbonylglycine were tion was stirred in the ice bath for 45 min and a t room temperatetrahydrofuran, dimethylformamide, and methylene chloride, ture overnight. The resulting suspension was filtered and the respectively. After the first reaction cycle, the unreacted 6residue was washed well with tetrahydrofuran. The filtrate was aminocaproyl resin was acetylated with acetic anhydride (3 ml) evaporated i n vacuo to yield a residue which was dissolved in 50 and triethylamine (1 ml) in 20 ml of dimethylformamide. Three ml of hot methanol. Upon cooling 1.OO g (32%) of white crystals syntheses of protected heptapeptide resin were carried out, all separated out of the solution: mp 160-162'; Rr (PW) 0.75, Rf starting with N-t-butyloxycarbonyl-c-aminocaproylresin (3.20 g, (RAWP) 0.75. Recrystallization from ethyl acetate did not 2.20 mmol). Amino acid analyses were performed after each raise the melting point. coupling step during the first synthesis of protected heptapeptide Anal. Calcd for Cl7Hl8N2O1: C, 64.95; H, 5.77; N, 8.91. resin. The results are given in Table 11. Found: C, 64.80; H, 5.92; N, 8.69. Glycyl-p-aminobenzyl Alcohol Hydrochloride ( 8 ).-The above TABLE Ir compound (0.385 g, 0.001 mol) was dissolved in 150 ml of 95% Cooling ethanol containing 1 ml of 1 N HC1. Palladium (5%) on charstep G~Y Bzim-His El3C coal (0.100 g) was added and the suspension was hydrogenated at 1 0.80 46 psi for 1.5 hr. The catalyst was removed by filtration and 1.0 the filtrate was evaporated i n vacuo. The resulting residue was 2 0.72 0.80 1.0 dissolved in hot methanol and ether was added until the solution 3 1.62 0.78 1.0 became faintly turbid. White needles separated from the solu4 2.05 0.76 1.0 tion upon cooling. The material slowly decomposes above 280': Rr (PW) 0.27, Rr (BAWP) 0.57. Glycyl-p-aminobenzoylglycyl-im-benzyl-~-histidyl-glycyl-pAnal. Calcd for CsHlaNzOzC1: C, 49.89; H, 6.05; N , 12.93; aminobenzoyl-e-aminocaproic Acid Dihydrobromide (4).-The C1,16.36. Found: C, 49.79; H,6.35; N, 12.87; C1,16.49. above protected heptapeptide resin was suspended in 40 ml of Reaction of 6 with Sodium Hydroxide.-A solution of 6 (0.130 trifluoroacetic acid and a stream of hydrogen bromide (pretreated g, 0.372 mmol) in 1.50 ml of methanol was treated with 0.50 ml with 10% resorcinol in acetic acid) was passed through the susof 1.50 N sodium hydroxide solution and the resulting solution pension for 30 min with occasional shaking. The mixture was was stirred a t room temperature for 29 hr. At 0.25 hr and 29 filtered and the resin was washed with trifluoroacetic acid (two hr, 0.050-ml aliquots were removed and treated with trifluoro20-ml portions), trifluoroacetic acid-methylene chloride (1 :I ) acetic acid (10 ml) for 0.50 hr. The solutions were evaporated (20 ml), and methylene chloride (two 25-ml portions). The i n vacuo. The resulting materials were dissolved in 1%HCI and pooled filtrates were evaporated i n vacuo to yield a brown oil. applied to an amino acid analyzer previously calibrated with The oil was taken up in methylene chloride which was removed glycine, glycyl-p-aminobenzoic acid," and compound 7. The i n vacuo. The process was repeated and the resulting oil was presence of 0.008 mmol of glycine (1,9%) and 0.322 mmol of 7 triturated with ether to give a cream-colored powder. The ma(87%) was shown for the 0.25-hr sample. In the 29-hr sample, terial was washed well with ether and dried in vacuo to give 1.24 g the release of 0.031 mmol of glycine (8.4%) and 0.340 mmol of 7 of crude material displaying one major component and several (92%) was indicated. No glycyl-p-aminobenzoic acid could be minor components in BAW, PW, and BAWP. Amino acid detected in either sample. analysis gave Gly:Bza*-His:Eac (3.18:1.00:1.37). The other Reaction of 7 with Sodium in Liquid Ammonia.-A solution of syntheses yielded 1.10 g (first synthesis) and 1.40 g (second syn7 (0.106 g, 0.422 mmol) in approximately 20 ml of liquid ammonia thesis) of crude heptapeptide dihydrobromide 4 . (freshly distilled from sodium) was treated with freshly cut pieces The crude material (1.16 g) was purified on a column (4.2 X of sodium until a blue color persisted for 30 sec. The color was 53 cm) of Sephadex LH-20 washed with methanol. Purified discharged with ammonium bromide and the ammonia was alheptapeptide dihydrobromide (4) (0.566 g) was obtained in 33% lowed to evaporate a t atmospheric pressure and then under reyield (based on N-t-butyloxycarbonyl-e-aminocaproylresin): duced pressure. The resulting residue was dissolved in 58 ml of mp 162-164'; [ c Y ] ~-7.50' ~ D (c 1, methanol); R f (BAW) 0.09, 50% aqueous acetic acid. An aliquot of this solution was placed Rr (PW) 0.31. on an amino acid analyzer previously calibrated with glycine, Anal. Calcdfor C89HlTN908Brz: N, 13.56. Found: N, 13.68. compound 7, and compound 8 . The chromatogram indicated Amino acid analysis gave Gly: Bzam-His:Eac(3.04: 1.00: 1.10). the release of 0.313 mmol of glycine (38% based on 0.844 mmol cyclo-(Glycyl-p-aminobenzoylglycyl-imbenzyl -I,- histidylglycylof glycine in 0.422 mmol of 7), 0.082 mmol of 8 (19%), and 0.004 p-aminobenzoyl-e-aminocaproyl) (3).-Heptapeptide dihydrommol of 7 (0.9%). bromide 4 (0.156 g, 0.168 mmol) was dissolved in 80% aqueous N-t-Butyloxycarbonyl-e-aminocaproylResin.-A solution of methanol (168 ml). The solution was cooled to 0' and N,N'N-t-butyloxycarbonyl-e-aminocaproicacid16 (5.39 g, 23.3 mmol) dicyclohexylcarbodiimide (0.346 g, 1.68 mmol) was added in and triethylamine (3.24 ml, 23.3 mmol) in 40 ml of ethanol was methanol (4 ml). The solution was stored a t 5' for 2 days and added to 10.00 g of chloromethylated polystyrene-2% divinyl a t room temperature for 13 days. Acetic acid (2 ml) was added benzene copolymer, 200-400 mesh (2.33 mmol of Cl/g). The mixto destroy unreacted N,N'-dicyclohexylcarbodiimide and the ture was stirred under reflux for 46 hr and filtered. The resin was solution was evaporated i n vacuo. The resulting mixture was washed thoroughly with ethanol and methanol. A sample of resolved on a column of Sephadex LH-20 (washed with methanol) the dried resin was hydrolyzed in 1: 1 dioxane-12 N HCI for 24 to yield 0.031 g (2501,) of a white material which began to brown hr a t 110'. Amino acid analysis showed the e-aminocaproic a t 245' and charred at 280-290". Thin layer chromatograms of acid content to be 0.620 mmol/g. N-t-Butyloxycarbonylglycyl-p-aminobenzoylglycyl-im-benzyl-the material gave single components that were ninhydrin and Pauly negative and chlorine positive: Rr (BAW) 0.19, Rr (PW) L-histidylglycyl-p-aminobenzoyl-e-aminocaproylResin.-A por0.54, Rt (BAWP) 0.60. tion of N-t-butyloxycarbonyl-e-aminocaproylresin (3.20 g, 2.20 A w l . Calcd for CagH&TeOT*CHaOH: C, 61.44; H , 6.06; mmol) was treated in the following manner for the incorporation N, 16.12. Found: C, 61.66; H, 5.87; N, 16.36. of each residue: (1) washed (three 50-mI portions) with methylene Amino acid analysis gave Gly:Bzam-His:Eac (2.88: 1.00:0.94). chloride, (2) mixed with trifluoroacetic acid-methylene chloride cycZo-(Glycyl-p-aminobenzoylglycyl-~-histidylglycyl-p-~ino(2:3) (50 ml) for 15 min, (3) washed (three 50-ml portions) with benzoyl-e-aminocaproyl) (2) (Directly from 4).-A solution of 4 methylene chloride, (4) washed (three 50-ml portions) with eth(0.465 g, 0.500 mmol) in 80% aqueous methanol was cooled to 0' anol, (5) washed (three 50-ml portions) with chloroform, (6) and N,N'-dicyclohexylcarbodiimide (1.03 g, 5 .OO mmol) in mixed with 40 ml of 10% triethylamine in chloroform for 10 min, methanol (20 ml) was added. The solution was stored a t 5' for 3 (7) washed (three 50-ml portions) with chloroform, (8) washed days and a t room temperature for 3 days. Acetic acid (6 ml) was (three 50-ml portions) with coupling solvent, (9) introduced 8.80 added and the solution was evaporated to near dryness in vacuo. mmol of appropriate N-t-butyloxycarbonyl amino acid or peptide The wet residue was suspended in 50% aqueous acetic acid (50 in 30-40 ml of coupling solvent and mixed for 10 min, (10) intro-
PHOSPHORYLATION OF UNPROTECTED RIBONUCLEOSIDES ml) and shaken well. The suspension was filtered and the residue was washed with additional aqueous acetic acid (50 ml) and filtered. Palladium (5%) on charcoal (0.500 g) was added to the combined filtrates. The suspension was hydrogenated (40 psi) a t room temperature for 66 hr. An aliquot taken at 66 hr gave a single Pauly-positive spot ( R f 0.41) on a thin layer plate developed with PW. The suspension was filtered and the filtrate was evaporated in vacuo to give 0.858 g of a foamy residue. A portion (0.644 g) of the residue was dissolved in 6 ml of butanolwater (6: 1) and applied to a column (2.3 x 74 om) of cellulose powder that had been washed with the same solvent mixture. Cyclic heptapeptide 2 was obtained as 0.057 g (23%) of a fine white material which charred a t 279-282”. Thin layer chromatograms gave single components that were Pauly and chlorine positGe and-ninhydrin negative: Rr (BAW) 0.09, Rr (PW) 0.39, Rr (BAWP) 0.47. -Anal. Calcd for C Z . Z H ~ ~ N ~ O ~ .C, H ~56.71; O: H, 5.79; N, 18.60. Found: C,56.42; H, 5.51; N, 18.51. Amino acid analysis gave Gly :His :Eac (2.84 :1.OO : 1.OO) . Compound 2 has very limited solubility in water and is sparingly soluble in most organic solvents (methanol, pyridine, dimethyl sulfoxide, dimethylformamide) which precluded a molecular weight determination via osmometry. Compound 2 is soluble in acidic solvents (trifluoroacetic acid, formic acid, and 50% aqueous acetic acid.) Compounds 2 and 3 were submitted for mass spectrometry. The samples were introduced using a direct probe. High-inlet temperatures ( >290’) were required to volatilize the materials. No parent peaks were observed as the peptides decomposed under these conditions. The largest observable fragments were approximately 500 mass units.27 X-Ray Determinations of 2.28-Compound 2 (2.2 mg) was dissolved in 50% acetic acid (0.22 ml) and added to a flask containing approximately 30 ml of 0.01 M phosphate buffer (pH 7.19). The flask was sealed and upon several weeks’ standing small (27) The authors are indebted to Mr. Karl Kohler (Department of Chemistry, Indiana University, Bloomington, Ind.) and to Dr. William Hargrove (Eli Lilly and Co., Indianapolis, Ind.) for several attempts to obtain the molecular weights of compounds a and 3 by mass spectrometry. (28) This work was performed by Dr. Jean Hamilton of this department.
J . Org. Chem., Vol. 36,No. 9, 1970 2881
clusters of very fine needles (barely visible without magnification) appeared. X-ray diffraction patterns were obtained with difficulty and the cell dimensions which were obtained from rotation and Weissenberg photographs are as follows: a = 9.32 f 0.03 A; b = 9.95 i 0.03 A; c = 36.57 f 0.02A. The crystal system is orthorhombic, but the space group was not determined owing to difficulties in obtaining good films. As the crystals were small and badly formed, it was impossible to measure the density with any reasonable accuracy. Assuming four molecules per unit cell and a minimum density of 1.30 g/cc, the molecular weight would be 665 (theory 678). The crystals exhibited no obvious effects of drying out, so it is unlikely that they contain much solvent. Kinetic Measurements.-The kinetic runs were performed with a Cary 14 recording spectrophotometer using a 10-cm silica cuvette. The reactions were carried out in 0.01 M phosphate buffer (pH 7.19) containing less than 0.4% organic solvents. Recrystallized imidazole was used for comparison with compound 2. The cuvette was filled with 50 ml of buffer, with or without catalyst, and placed in a 26 f 0.1’ circulating bath for at least 15 min. The cuvette was placed in the cell compartment (thermostated a t 26’) and balanced against air. Tine substrate, 2,4-dinitrophenyl acetate, was added in acetonitrile and the cuvette was gently agitated and returned to the cell compartment. The recording of a run began 60 sec after the addition of substrate. The appearance of 2,4-dinitrophenylate anion was measured at 360 mp. All reactions were followed to greater than 90% completion. At the end of each run the pH was 7.19 f .02. Infinity absorbances were taken at greater than ten halflives. First-order rate plots were obtained for all reactions. The first-order rate constants were obtained by the method of Guggenheim.29 The kinetic results are summarized in Table I.
Registry No.-2, 25383-41-9; 3, 25533-69-1; 4, 25442-38-0; 5 , 25442-39-1 ; 6, 25383-67-9; 7, 2538368-0; 8, 25383-69-1; N-t-butyloxycarbonylglycyl-paminobenzoylglycine, 25333-70-4; N-carbobenzoxyglycyl-p-aminobenzyl alcohol, 25383-71-5. (29) “Kinetics and Mechanism,” 2nd ed, Frost and Pearson, Ed., Wiley, New York, N. Y.. p 49,1961.
Selective Phosphorylation of the cis-2’,3’-Diol of Unprotected Ribonucleosides with Trimetaphosphate in Aqueous Solution R. SAFFHILL The Salk Institute, San Diego, California 9,9112 Received February 26, 1970 Unprotected ribonucleosides are selectively phosphorylated at the cis-’2’,3’-diol in high yield by trimetaphosphate at high pH. The reaction is used to prepare several ribonucleoside 2’(3’)-phosphates including wcytidine 2’(3’)-phosphate.
Most methods of phosphorylating unprotected ribonucleosides with activated phosphates or with orthophosphate and a condensing agent yield mixtures of 2‘-, 3’-, and 5‘-monophosphates as well as di- and triphosp h a t e ~ . ’ - ~ The 5’-phosphate is usually the major monophosphate formed, atlhough the ratio of the products does depend on the nature of the reactants and ~ o l v e n t . Help ~ and Smrt5 have reported the synthesis of ribonucleoside 2’(3’)-phosphites from the unprotected ribonucleoside and triethyl phosphite and their oxidative cyclization to 2’,3‘-cyclic phosphates, but there is no convenient method of directly phosphorylating the (1) G . R . Barkerand J. M . Gulland, J . Chem. SOC.,231 (1942). (2) J. Baddily, J. G. Buchanan, and R. Letters, ibid., 1000 (1958). (3) A. Beck, R. Lohrmann, and L. E. Orgel, Science, 167,952 (1967). (4) G.R.Barker and G. E . Foll, J . Chem. SOC.,3798 (1957). (5) (a) A. Holg and J . Smrt, Collect. Czech. Chem. Commun., S I , 1528 (1966); (b) A . Holj. and J. Smrt, ibid., 31, 1562 (1967).
cis-2’,3‘-diol of an unprotected ribonucleoside in good yield. Feldmane has reported that sodium trimetaphosphate reacts with ethylene glycol at high pH to give P-hydroxyethyl phosphate. Sucrose yields sucrose phosphate under similar conditions, although the position of phosphorylation was not determined. Here we wish to report the synthesis of ribonucleoside and ribonucleotide 2’(3’)-phosphates by a modification of this reaction. I n a preliminary experiment adenosine (Ia) was treated with 10 mol equiv of sodium trimetaphosphate and 10 mol equiv of 1 N aqueous sodium hydroxide; there was a 63% conversion to adenosine 2’(3’)-phosphate (IIa) on standing overnight at room temperature. There was no further reaction after an additional day. When tri(tetramethy1ammonium) trimetaphosphate (6) W. Feldman, Chem. Ber., 100,3850 (1967).
